Chemometrics for Discriminating the Geographical Origin of Korean and Imported Wheat Flour

Abstract

Consumers consider geographical origin an important factor when choosing a product. Therefore, research is continuously conducted to determine the country of origin of products such as wheat. In this study, inorganic elements were used to identify Korean and imported wheat flour using inductively coupled plasma (ICP) analysis. Chemometrics was employed to suggest a method for determining the country of origin, which included principal component analysis (PCA), partial least squares discriminant analysis (PLS-DA), and orthogonal partial least squares discriminant analysis (OPLS-DA). In PLS-DA and OPLS-DA, 21 and 19 elements were investigated to influence the determination of geographical origin. The elements with the greatest discriminatory power were ⁸⁸Sr and Fe. OPLS-DA had the highest Q² and R² scores, clearly differentiating Korean and imported wheat flour. In a heatmap, the inorganic element content of each flour group was observed, and canonical discriminant analysis (CDA) confirmed the country of origin with 100% accuracy. CDA showed 98% accuracy with only one imported sample error in the validation set. Furthermore, this work can inform method selection for the country of origin determination of other product types.

Keywords: CDA; ICP analysis; OPLS‐DA; chemometrics; geographical origin; heatmap.

FIGURE 1

Discriminant analysis of Korean and imported wheat flour based on inorganic elemental content data obtained from ICP analysis. (A) PCA loading plot and (B) PCA plot; figure legend: Korean (blue) and imported (red) wheat flour.

FIGURE 2

PLS‐DA (Partial Least Squares Discriminant Analysis) results show differentiation between Korean and imported wheat flour samples. (A) The loading plot reveals a clear separation of elemental variables, with most key contributors (e.g., Fe, Mn, P, K) positioned toward the Korean sample side, indicating their higher association with domestic products. (B) The score plot demonstrates improved group separation compared to PCA, with reduced overlap and clearer classification. (C) VIP (Variable Importance in Projection) analysis identifies 21 elements (VIP ≥ 1) that significantly contribute to origin discrimination, many of which are enriched in Korean samples.

FIGURE 3

OPLS‐DA (Orthogonal Partial Least Squares Discriminant Analysis) results demonstrate clear discrimination between Korean and imported wheat flour samples based on elemental composition. (A) The loading plot shows that most variables of importance (VIP ≥ 1) are skewed to the left, suggesting their strong association with Korean samples. (B) The score plot indicates improved class separation compared to PCA and PLS‐DA, with minimal overlap between groups. (C) VIP analysis identified 19 elements (e.g., Fe, Mn, P, Mg, Ba) that significantly contribute to origin classification. (D) The ROC curve shows perfect classification performance (AUC = 1), supporting the model's discriminative power. (E) A permutation test (n = 200) yielded R ² = (0, 0.278) and Q ² = (0, −0.705), validating the robustness and predictive reliability of the OPLS‐DA model.

FIGURE 4

Heatmap visualization for determining the geographical origin of wheat flour based on the content of 53 inorganic elements detected by ICP analysis. Rows represent individual samples (blue: Korean wheat flour; red: imported wheat flour), and columns represent elements. Element concentrations are shown from low (dark blue) to high (dark red). A clear pattern emerges in the lower group of elements (Na–Zn), where Korean samples generally exhibit higher concentrations than imported samples. Notably, most of the VIP elements identified by OPLS‐DA fall within this group, reinforcing their discriminative power. In contrast, the upper group of elements shows minimal variation between origins and contains only a few VIPs, further highlighting the relevance of the lower group in geographical origin classification.